Department of Systems & Computational Biology

Machine Learning Reading Group Meetings


 As you may be aware, machine learning (ML) is a rapidly developing branch of artificial intelligence (AI), with a long history rooted in statistics, computer science, and physics. Its recent successes in AI applications—including automated-image understanding, ads and TV show recommendations, flying an unmanned helicopter upside down, and beating a grand-master in a Go tournament—have featured prominently in the scientific and popular news. The excitement for ML is also growing in the biomedical community as it becomes clear that ML could assist and improve practical applications ranging from medical image analysis to discovering patterns in large patient databases, as well as to address basic science questions such as interpreting genetic networks and explaining the function of brain circuits.    

 In August 2016, faculty of systems and computational biology started the Reading Group on Recent Advances in Machine Learning, an informal, monthly meeting in which we discuss the newest publications and techniques in ML. The meeting offers the opportunity to discover new applications of ML and learn the techniques that make such advances possible. We will go into some mathematical and computational details, while also discussing higher-level conceptual issues.    

 In each meeting, a volunteer chooses a paper(s) and is in charge of presenting it with slides. The volunteer shares the paper(s) at least a week in advance so everyone interested in attending has time to read and come prepared with questions/comments. The meeting typically lasts 1 to 1.5 hours, with slide presentation, questions and discussions. Everyone is welcome to attend and join the interactive discussion. Please see below to view the full calendar and meeting locations (which can vary month to month), and contact Dr. Ruben Coen-Cagli at ruben.coen-cagli@einsteinmed.org if you would like to present at a future meeting (We have openings available!). Also, be sure to bookmark this page for easy reference and updates.  

   

        

 Calendar for 2019-20  

    

   

 OCT 28 (2pm, Price 451) Ruben Coen-Cagli (Einstein) – Neural population control using ‘mind-blowing’ synthetic images.  

 NOV 18 (noon, Price 551)  Saad Kahn (Einstein, Kelly lab) – Interpretation methods for deep learning models: saliency mappings 

 DEC 16 (noon, Price 551) Sacha Sokoloski (Einstein, Coen-Cagli lab) – State-of-the-Art of Artificial General Intelligence.  

 JAN 27 (noon, Price 551) videolecture by Surya Ganjuli - Deep Learning Theory: From Generalization to the Brain  

 FEB 24 Aude Genevay (MIT, Geometric Data Processing Group) - Optimal transport and applications. 

 APR 20 special Zoom-only session on AI/ML initiatives for COVID-19. Panelists: Aviv Bergman, Libusha Kelly, Saad Kahn, Jonathan Vacher, Dylan Festa, Sacha Sokoloski, Yehonatan Sella, Ruben Coen-Cagli. Zoom meeting ID 999-4769-0203 at noon. A list of resources is available here. 

 MAY 18 (noon, Zoom only) Niko Kriegeskorte (Columbia University, Director of Cognitive Imaging) - Testing deep neural network models of human vision with brain and behavioral data.  Zoom Meeting (Meeting ID 859 907 5850 -- Password: MLreading)  

    

 Past Meetings  

 2019 May. Rajesh Ranganath (NYU Courant Institute)  

 2019 March. Multiscale interpretable models of neural dynamics, presented by Memming Park (Stony Brook)  

 2019 February. Probabilistic segmentation with U-NET, presented by Ruben Coen-Cagli (Einstein)  

 2019 January. Enhancing fluorescence microscopy with deep learning, presented by Adrian Jacobo (Rockefeller)  

 2018 November. Geometric deep learning, presented by Saad Kahn (Einstein)  

 2018 October. Adversarial networks, presented by Ruben Coen-Cagli (Einstein)  

 2018 May. Multiscale Methods for Networks, presented by Bo Wang      

 2018 April. The scattering transform, presented by Jonathan Vacher      

 2018 March. Topological data analysis, presented by Michoel Snow      

 2018 February. Recurrent Neural Networks for Sequence Learning, presented by Sacha Sokoloski      

 2018 January. Opening the black box of Deep Neural Networks via Information, presented by Saad Khan      

 2017 December. Probabilistic programming with STAN, presented by Dylan Festa      

 2017 November. Visualizing Data using t-SNE, presented by Daniel Pique      

 2017 October. Deep Convolutional Neural Networks, presented by Sacha Sokoloski      

 2017 September. Bayesian sparse priors and shrinkage, presented by Shuonan Chen    

 

 2017 August. The Variational Autoencoder, presented by Ruben Coen-Cagli    

 

Click here to log in